JPH0460389B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Industrial Application) The present invention relates to an image enlarging/reducing method for at least one of enlarging or reducing an image.

(Prior Art) As an image enlargement/reduction method for enlarging or reducing an image using a pixel density conversion method, the simple thinning method, SPC method, logical sum method, 9-division method, projection method, etc. are conventionally known.

(Problems to be Solved by the Invention) Such a method is used in printers, facsimiles, etc., and is mainly known as a method for enlarging/reducing line drawings and character drawings. When these methods are applied to an image with gradation, the magnification is an integer multiple, and the dot pitch of the image after scaling is usually different. Furthermore, when multiplying by a rational number, there is a problem that image deterioration (moiré fringes) occurs, resulting in an image that is not suitable for practical use.

For example, images that have been binarized using a dither matrix have different characteristics depending on the dither matrix used, and depending on the type of dither matrix used, the image There are some properties that are not suitable for enlarging or reducing.
In particular, reduction processing is performed on images that have been binarized using a dot-distributed dither matrix.
When enlarging an image that has been binarized using a concentrated dither matrix, the deterioration in image quality becomes noticeable.

The present invention has been made in view of the above-mentioned problems, and its purpose is to make it possible to keep the mesh pitch unchanged before and after enlargement/reduction, and to be able to perform enlargement/reduction by a rational number of times.
The object of the present invention is to provide an image enlargement/reduction method that does not cause image deterioration in that case.

(Means for Solving the Problems) The present invention for solving the above problems divides a first binarized image, which is an original image, into unit regions each consisting of a plurality of pixels, and a predetermined number of adjacent units of the The area is divided into divided areas, and the variation in the number of black pixels within the unit area is determined for each divided area. For the divided area where the variation is small, a large-sized threshold matrix is used to calculate the number of black pixels within the unit area. For divided regions with large variations, a second binarized image is obtained using a small-sized threshold matrix, and the second binarized image is
The valued image is divided into multiple blocks f ₁₁ , f ₁₂ , ..., f _1j , ..., f ₂₁ , f ₂₂ , ..., consisting of n×n pixels.
When dividing into f _2j , ..., f _ij , ... and obtaining an enlarged/reduced image when the enlargement/reduction magnification is m/n, the enlarged/reduced image is divided into Blocks F ₁₁ , F ₁₂ , . . . are arranged in the same number as the blocks of the second binarized image.
..., F _1j , ..., F ₂₁ , F ₂₂ , ..., F _2j , ...,
F _ij , ..., and each block consists of m×m pixels, and only the density matrix pattern of the block f _ij in the i-th row and j column of the second binarized image is two-dimensionally Assuming a mother pattern arranged in an array, a pattern consisting of m×m pixels is cut out from the mother pattern, and a density matrix pattern of the block F _ij in the i-th row and j column of the enlarged/reduced image is obtained. This feature is characterized in that enlarged/reduced images can be obtained by using the image as an image.

(Example) Hereinafter, the image enlargement/reduction method of the present invention will be specifically explained using FIGS. 1 to 7.

Here, the binary image to be enlarged/reduced is usually one obtained using a dither method, a density pattern method, a halftone method, or the like. In the method of the present invention illustrated in FIG. 1, a dot-distributed (Bayer-type) dither matrix with a size of 4×4 (see FIG. 2 A) is used as the threshold matrix, and the original image A is
(See Figure 2 B) is binarized as shown in Figure 2 C.
Assume that a digitized image B is given as an object to be enlarged or reduced. In FIG. 2, the numbers in the dither matrix and in the original image A indicate the standardized density levels, and the numbers in the dither matrix and the original image A indicate the standardized density levels,
The pixels in the shaded area are black pixels.

First, in step, the first binarized image, which is the original image, is divided into unit areas UA of appropriate size. Here, it is divided into the same size as the dither matrix, that is, into 4×4 sizes as shown in FIG. 2C. Then, the number of black pixels (or the number of white pixels) in each unit area UA is counted (step).

Next, a partitioned area MA is formed by a plurality of adjacent unit areas UA (in this example, a matrix of 2 x 2 vertically x horizontally), and for each partitioned area MA, the variation in the number of black pixels within the unit area UA is Investigate (step). The value of this variation ε is, for example,
Calculate the average number of black pixels in the four unit areas UA that form the divided area MA, and subtract this average number of black pixels from each of the four unit areas, and find the largest deviation (absolute value). Use. and,
When this variation ε is less than a preset reference value ε ₀ , a large-sized threshold matrix is used to create a density matrix pattern (white,
If the variation ε is larger than the reference value ε ₀ , the density matrix pattern of the corresponding divided area MA is determined using a small-sized threshold matrix. To give a concrete example, the numbers in Figure 3A are the number of black pixels in each unit area UA, and if the variation ε in the left partitioned area MA (MA ₁ ) is determined as described above, ε = 0.5, and the right side The variation ε in the divided area MA (MA ₂ ) is found to be ε=2.5. Here, if the reference value ε ₀ is set to ε ₀ =1.5, for example, ε<ε ₀ in the divided area MA ₁ and ε>ε ₀ in the divided area MA ₂ . Therefore,
In this embodiment, a large-sized threshold matrix is used for the divided area _MA1 , and a small-sized threshold matrix is used for the divided area _MA2 to obtain the density matrix pattern. For example, partition area MA ₁
For , the size corresponding to the partitioned area MA (8
×8) using the threshold matrix,
For MA ₂ , a density matrix pattern is determined using a threshold matrix of a size (4×4) corresponding to the unit area UA. Figure 4 (a) shows each divided area MA based on the number of black pixels within the area divided by the size of the threshold matrix used (see Figure 3 (b)).
The density matrix pattern is shown above.
The determination of the density matrix pattern in this example is
In the case of an 8 x 8 size threshold matrix, use the dither matrix shown in Figure 5A, and calculate the relationship between this dither matrix and the number of black pixels (corresponding to the density level) in each section. This is done by comparison, and in the case of 4×4 size, the fifth
This is done by comparing the dither matrix shown in Figure B with the number of black pixels in each section. For example, in the case of section K, the number of black pixels is 10, so in the dither matrix shown in Figure 5B, the portion with a density level of 10 or less becomes a black pixel, and the density level as shown in Figure 4B described above becomes a black pixel. It becomes a matrix pattern. A second binarized image is obtained from the density matrix pattern thus obtained (step).

Next, this second binarized image is divided into a plurality of blocks f ₁₁ , f ₁₂ , ..., f _1j , ... consisting of n×n pixels.
..., f ₂₁ , f ₂₂ , ..., f _2j , ..., f _ij , .... For example, as shown in FIG. 4B, it is divided into 4×4 blocks BK (step). below,
The density matrix pattern of this block is
This is called the density matrix pattern.

Here, blocks F ₁₁ are arranged so that when the magnification/reduction magnification is m/n, the number of enlarged/reduced images is the same as the blocks of the second binarized image in both the row direction and the column direction. ,F ₁₂ ,...,F _1j ,...,
Consisting of F ₂₁ , F ₂₂ , ..., F _2j , ..., F _ij , ...,
It is assumed that each block consists of m×m pixels, and only the first density matrix pattern of the block f _ij in the i-th row and j column of the second binarized image is two-dimensionally A pattern configured by arranging the patterns (herein referred to as a mother pattern) is assumed. Next, a pattern consisting of m×m pixels is cut out from this mother pattern, and a density _matrix pattern (hereinafter, this is called the second It is used as a density matrix pattern (called a density matrix pattern). A second density matrix pattern is obtained for each block, and the patterns are arranged in block order to obtain an enlarged/reduced image (step). FIG. 6A is an enlarged image obtained in this way with an enlargement factor of 5/4, and FIG. 6B is a reduced image with a reduction factor of 3/4. Here, as described above, the size ratio of the first and second density matrix patterns depends on the enlargement/reduction magnification, and the vertical and horizontal sizes of the second density matrix pattern are the same as those of the first density matrix pattern. This is the vertical and horizontal size of the density matrix pattern multiplied by the vertical and horizontal expansion/reduction magnification. Therefore, in the example of FIG. 6 where the vertical and horizontal magnifications are equal, the size in FIG. 6A is 5×5, and the size in FIG. 6B is 3×3. FIG. 7 is an explanatory diagram (magnification: 5/4) showing an example of obtaining a second density matrix pattern from a first density matrix pattern. Block f ₁₁ , f ₁₂ , f ₁₃ ,
The plane consisting of ..., f ₂₁ , f ₂₂ , ..., ... is the second
density matrix pattern (size 5x5)
Blocks F ₁₁ , F ₁₂ , F ₁₃ , ..., F ₂₁ , F ₂₂ , ...
. . . (The pixel in the first row and first column of block _f11 and the pixel in the first row and first column of block _F11 are in a positional relationship such that they overlap). First, as a second density matrix pattern corresponding to block F ₁₁ , all blocks f ₁₁ ,
From the mother pattern assuming that f ₁₂ , f ₁₃ , ..., f ₂₁ , f ₂₂ , ..., ... contain the same first density matrix pattern corresponding to block f ₁₁ , block F Use the pattern cut out at position ₁₁ . Similarly, the second density matrix pattern corresponding to block _F12 is as follows:
All blocks f ₁₁ , f ₁₂ , f ₁₃ , ..., f ₂₁ , f ₂₂ , ...,
..., a pattern at the position of block _F12 is cut out from a mother pattern which is assumed to contain the same first density matrix pattern corresponding to block _f12 . That is, block
The second density matrix pattern corresponding to F _ij includes all blocks f ₁₁ , f ₁₂ , f ₁₃ , ..., f ₂₁ ,
A pattern at the position of block F _ij is cut out from a mother pattern that is assumed to contain the same first density matrix pattern corresponding to block f _ij in f ₂₂ , . . . , . In the case of expansion, blocks f ₁₁ , f ₁₂ , ..., f _1j , ..., f ₂₁ ,
When a mother pattern is constructed from only f ₂₂ , ..., f _2j , ..., f _ij , ..., the mother pattern area is
F ₁₁ , F ₁₂ , ..., F _1j , ..., F ₂₁ , F ₂₂ , ...,
Since the area is smaller than the area consisting of F _{2j ,} . There is a need.

In addition, in the above step, if the corrected black pixels obtained by using a spatial filter that takes into account the number of black pixels in the surrounding unit area, etc. are used instead of using the number of black pixels obtained in the step as is, image enhancement can be improved. (filtering) can be performed at the same time. Furthermore, if the number of black pixels obtained in the step is converted to another number of black pixels based on a predetermined gradation conversion curve, gradation processing becomes possible. The type of gradation processing to be performed is determined by the gradation conversion curve (the horizontal axis is the original black pixel number and the vertical axis is the converted black pixel number).

Furthermore, in the method of the present invention, in the above step, a part of the pattern of the binarized image (corresponding to FIG. 2) obtained in the step is used as it is as the first density matrix pattern (small size). You can also do it.

Further, a specific method for obtaining the pattern of each block after scaling described above may be to actually configure the above-mentioned mother pattern in memory and read data at a predetermined address from it to obtain the pattern. , this would require an extremely large memory capacity. Therefore, instead of actually creating a mother pattern, attention may be paid to the periodicity of the pattern, and the density of each pixel may be determined in the following manner to obtain a pattern after scaling.

That is, in the case of enlargement (magnification m/n), the x-th pattern in the I-th block in the row direction of the new block (block after enlargement) is the same as the pattern in the I-th block in the row direction of the block before enlargement. AD1 = mod [x + mod [(I-1) (m-n), n] + 1-2, n] + 1st row pattern, and the y column in the J-th block in the column direction of the new block. The eye pattern is equal to the pattern of AD2=mod[y+mod[(J-1)(m-n), n]+1-2,n]+1st column in the Jth block in the column direction of the block before expansion. , on the other hand, reduction (magnification m/
n), the x-th row pattern in the I-th block in the row direction of the new block (block after reduction) is AD1=mod[x+mod[( I-1) (n-|n -m|),n]+1-2,n]+1 The pattern in the y-th column in the J-th block in the column direction of the new block is It is equal to the pattern of AD2=mod[y+mod[(J-1)(n-|n-m|),n]+1-2,n]+1st column in the Jth block in the column direction of the block before reduction.

Here, mod [p, q] is the remainder of p÷q,
Of course it is smaller than q.

Therefore, if the pixel density of each block before scaling is stored in memory, it can be easily updated by reading out the pixel density in the corresponding block before scaling using AD1 and AD2 as addresses in the row and column directions. You can get various block patterns. Naturally,
In order to obtain a pattern of equal size, the pixel density data of each block stored in this memory may be read out and used as is.

According to the above-mentioned enlargement/reduction method, no matter what kind of dither matrix is used to perform the binarization processing on the original binary image, the image has characteristics suitable for enlargement or reduction. Since the image is converted into a binary image and then enlarged or reduced, a high-quality enlarged/reduced image can be obtained. In particular, threshold matrices of multiple sizes are prepared, and in partitioned areas where density changes are gradual, gradation expression is prioritized over resolution by using large-sized threshold matrices, and conversely, in partitioned areas where density changes are rapid, small-sized threshold matrices are used. Since a threshold matrix is used to give priority to resolution, both gradation and resolution are excellent, and no detail is lost in the enlarged image.

FIG. 8 is a diagram showing an example of an apparatus for realizing the method of the present invention. In the figure, reference numeral 1 reads the first binarized image stored in the memory 2, divides it into unit areas each consisting of a plurality of pixels, and forms divided areas with a predetermined number of adjacent unit areas. A black pixel counting means counts the number of black pixels in a unit area for each partitioned area and stores the counted value in the memory 3; 4 reads the number of black pixels from the memory 3 and calculates the number of black pixels in the unit area for each partitioned area; A large-sized threshold matrix is used for divided regions with small variations, and a small-sized threshold matrix is used for divided regions with large variations in the number of black pixels within a unit area. A binarized image re-creation means 6 obtains a binarized image and stores it in the memory 5, and 6 reads out the second binarized image from the memory 5 and creates a first binarized image corresponding to the number of black pixels for each block. A first density matrix pattern creating means 8 reads out the first density matrix pattern of each block from the memory 7 and stores it in the memory 7, and a second density matrix pattern 8 of a size corresponding to the magnification is read out from the memory 7. The second density matrix pattern is converted into a density matrix pattern of
This is a density matrix pattern creation means. This second density matrix pattern creation means 8
stores the second density matrix pattern in all blocks in the memory 9. In this way, the memory 9 stores a binary image reconstructed using the second density matrix pattern of each block, that is, an enlarged or reduced binary image. Incidentally, if the plurality of memories mentioned above have a large memory capacity, a single memory is sufficient.

(Effects of the Invention) As explained above, in the present invention, a second density matrix pattern having a size corresponding to the magnification is obtained for each block from the first density matrix pattern. is unchanged,
Unlike the conventional method of obtaining enlarged/reduced images by changing the sampling interval, moiré fringes are less likely to occur. In addition, we prepare density matrix patterns of multiple sizes, and in divided areas where density changes are gradual,
A large-sized threshold matrix is used to give priority to gradation expression over resolution, and a small-sized threshold matrix is used to give priority to resolution in partitioned areas where density changes are large, so gradation and resolution are Both are excellent, and no detail is lost in the enlarged image.

[Brief explanation of the drawing]

FIG. 1 is a flowchart showing an example of the method of the present invention, FIG. 2 is an explanatory diagram showing an example of binarization, FIG. 3 is an explanatory diagram of divided areas, FIG. 4 is an explanatory diagram of block division, and FIG. The figure is an explanatory diagram of the dither matrix for obtaining the first density matrix pattern.
The figure is an explanatory diagram of the second density matrix pattern (enlarged/reduced image), FIG. 7 is an explanatory diagram of the method of obtaining the second density matrix pattern from the first density matrix pattern, and FIG. FIG. 2 is an explanatory diagram of an apparatus for carrying out the invention method. A...Original image, B...Binarized image,
UA...unit area, MA...divided area, f ₁₁ , f ₁₂ ,
..., f ₂₁ , f ₂₂ , ..., ... block of the first density matrix pattern, F ₁₁ , F ₁₂ , ..., F ₂₁ ,
F ₂₂ , ..., ... second density matrix pattern block.

Claims (1)

[Scope of Claims] 1. A first binarized image, which is an original image, is divided into unit areas each consisting of a plurality of pixels, a predetermined number of adjacent unit areas form divided areas, and each divided area is Find the variation in the number of black pixels within the unit area,
A large-sized threshold matrix is used for the partitioned areas with small variations, and a small-sized threshold matrix is used for the partitioned areas with large variations in the number of black pixels within the unit area to generate the second binarized image. The second binarized image is divided into a plurality of blocks f ₁₁ , f ₁₂ , ..., f _1j , ..., consisting of n×n pixels.
When dividing into f ₂₁ , f ₂₂ , ..., f _2j , ..., f _ij , ... and obtaining the enlarged/reduced image when the enlargement/reduction magnification is m/n, the enlarged/reduced image but,
Blocks arranged in the same number as the blocks of the second binarized image in both the row and column directions.
F ₁₁ , F ₁₂ , ..., F _1j , ..., F ₂₁ , F ₂₂ , ...,
It consists of F _2j , ..., F _ij , ..., and each block is m
Assume that the block consists of ×m pixels, and the block in the i-th row and j column of the second binarized image
Assuming a mother pattern consisting of only density matrix patterns of f _ij arranged two-dimensionally, a pattern consisting of m×m pixels is cut out from the mother pattern and is An image enlargement/reduction method characterized in that an enlarged/reduced image is obtained by using the block F _ij in the i-th row and the j-th column as a density matrix pattern.